XML: good, but no magic bullet (was Re: Perl is worse!)

[Alex Martelli]

"Steve Lamb" <grey at despair.rpglink.com> wrote in message
news:slrn8o6m3f.f91.grey at teleute.rpglink.com...
> On Sat, 29 Jul 2000 12:13:16 +0200, Alex Martelli <alex at magenta.com>
wrote:
> >(XML helps.  But don't anybody think it's anywhere close to
> >the panacea it's apparently considered to be in some circles;
> >that is just Yet Another recurrence of the never-ending dream
> >for the Magic Bullet:-).
>
>     Someone tell that to the people who wrote O'Reilly's Java & XML and,
hmm,
> there was another one out there.  If you want a good chuckle pick it up
and
> read the first two paragraphs of the preface and introduction.  In a
nutshell

Yeah, there's lot of hype about it, as is typically the case with so
many new technologies in our field.  In this case, it's a good and
very useful technology, and I, for one, find the hype particularly
irritating in cases like this -- because it raises expectations to
unreasonable levels, thus ensuring eventual backlash.

The point is: just as exceptions help you handle errors well, just
as locales (and facets) help you do internationalization right, so
does XML help you achieve interoperability (particularly exchange
of data) between heterogeneous programs.  But HELP is all each of
these good technologies can do: you must still understand what it
is you're doing, you must still do it properly.  It IS no doubt
easier to kill a bear with a gun than with a pocket knife, but if
your aim is totally off then even the gun can't help you enough...


>     Me, I filed the book away until I actually have a need for it in spite
of
> the author's view of its use in the world.  So far I have seen two
/examples/
> of its use at one conference and to me it looks like a markup language
with
> definable tags.

Yes, and a rather well-designed standard, with useful standardized
API's (DOM, SAX) and lots of promising extensions and applications
(although some of them bid fair to become too complicated for really
practical use, sigh -- and also, there's so MANY of them one risks
losing track:-).  The point is, it will not *by itself* solve your
interoperability problems, not even, by itself, all of your data
interchange problems; you must still apply it correctly, and with
clear understanding of its strengths, weaknesses, uses, limits.
Basically, like any other technology -- some more, others less.

Getting acquainted with the XML scene is not a bad idea, but IMHO
the best way to do that right now would be to study the annotated
standard (VERY good annotations!) then start playing around a bit
with Python's XML support (due to grow a lot in the near future).
I've got a dozen books on it, but none has given me much added
value wrt the annotated standard plus some practical experiments!-)


Alex